Neuroscience has upended the old myth that perfect pitch is a "you've either got it or you don't" gift. Modern training studies consistently show that adults can learn to name notes without a reference tone. Yet whenever these studies appear, skeptics are quick to note that each experiment trains only a handful of volunteers. Why are the sample sizes so tiny? And what happens when you scale that training to thousands of people?
Small Studies, Big Questions
To understand the criticism, it helps to look at the numbers. A 2019 training study from the University of Chicago required participants to complete three different exercises four times a week for eight weeks, roughly 32 hours of work. Though a full one third of the participants finished the program with near-perfect accuracy across all timbres and octaves, and all others showed consistent improvements above chance, the study's sample size was only six people. Even the largest investigations carry minimal statistical power. Another intensive training study published in 2020 also had participants meet conventional criteria for absolute pitch. The study's pool was limited to just 43 participants, so the 14% who learned amounted, again, to merely six people.
The average sample size for historical adult training studies is about twenty participants. Small sample sizes are typical of cognitive neuroscience, where fMRI and behavioral experiments have a median sample size of 28 participants. The common criticism is observably correct: historical adult training studies are too small to be statistically powerful. But researchers weren't aiming for statistical power. Absolute pitch has long been considered unlearnable, so their charter was to demonstrate that it can be learned at all.
Still, Why So Few Participants?
There are practical reasons behind tiny cohorts. Training someone to label pitches is genuinely demanding work, and the burden falls on both sides of a study. Most studies require hours of structured practice per participant, while researchers control and monitor conditions that make the data usable: instrument timbre, octave range, and response-time limits. They also have to make sure participants practice correctly and effectively, and that no one undermines the results, whether accidentally through faulty technique or deliberately by gaming the tasks. Adding participants multiplies the effort, so recruiting and supervising quickly becomes a limiting factor. With traditional methods, the work turns prohibitively labor-intensive well before a cohort reaches a hundred people.
This is not unique to absolute pitch research. Small cohorts draw the same scrutiny across neuroscience, where an editorial in The Journal of Neuroscience identifies low statistical power, frequently caused by small sample sizes, as a leading source of reproducibility problems. But the deeper constraint is money, and a 2018 analysis of how small samples undermine fMRI replicability put it directly:
Even relatively small studies can cost several tens of thousands of dollars, and the funding system throughout much of the world is not generally set up to enable the routine collection of large (e.g., N > 100) samples.
Turner et al., Communications Biology, 2018
At Scale
Small sample sizes aren't the only limitation. Participant pools are often made up of students whose availability follows the academic calendar, so training experiments tend to end when the semester does. That makes long-term outcomes, such as whether trained accuracy holds or fades over months and years, hard to capture and harder to publish, since following dozens of individuals over that span demands more resources and continuity than most labs can spare.
In contrast to the heavily constrained environments of published academic research, HarmoniQ has served as a platform for thousands of learners, showing the acquisition of absolute pitch on a much larger scale. The app has processed many millions of individual training trials, sufficient to map the progression of learning as a function of time and training volume, revealing consistent patterns in how users acquire pitch-labeling accuracy.
The platform's reach has also drawn interest from the research community: Stephen Van Hedger, lead author of the 2019 study that overturned the long-standing belief that absolute pitch cannot be learned in adulthood, and HarmoniQ are exploring a potential research collaboration. At this scale, the data confirms that the improvements seen in laboratory settings are not isolated flukes, but rather the expected outcome of structured, consistent training.